Can I vibe code Otter.ai?
AI meeting transcription with speaker labels and summaries.
Post-meeting transcription is a solved one-sitting build: record, run Whisper-class speech-to-text, summarize with an LLM. The gap is live: real-time transcription while the meeting happens, reliable speaker diarization, and auto-joining your calendar's calls as a bot. If you're happy pressing record and reading the summary five minutes after, you don't need the subscription.
Build a meeting transcription tool. Two inputs: upload an audio file, or record in the browser. Pipeline: transcribe with a speech-to-text API (word timestamps), then run an LLM pass that produces a 5-bullet summary, decisions, action items with owners, and rough speaker sections. Store meetings in a database with full-text search across transcripts. UI: meetings list, meeting page with summary on top and expandable transcript below (timestamps clickable to seek the audio), and one-click copy of the summary formatted for Slack. Next.js, object storage for audio, deployable to Vercel.
A drop-an-audio-file (or record-in-browser) tool that returns a transcript, speaker-guessed sections, action items, and a summary — archived and searchable.
- Live transcription during the meeting
- A bot that auto-joins Zoom/Meet from your calendar
- Accurate named speaker diarization
- Team workspaces and cross-meeting search polish
The auto-join bot means zero behavior change: they never press record, and every meeting just appears transcribed. Habit-free capture is the product.
A prompt is the first move, not the whole game. Course 01 teaches you to take a prompt like this one to a shipped, working app — reviewing, correcting, and steering the agent the whole way.
Start Course 01 →